On Tuesday 18 February 2014 17:11:43 Sachin Kamat wrote: > On 18 February 2014 16:33, Arnd Bergmann <arnd@xxxxxxxx> wrote: > > On Tuesday 18 February 2014 16:27:54 Rahul Sharma wrote: > >> > >> +static struct map_desc exynos5260_iodesc[] __initdata = { > >> + { > >> + .virtual = (unsigned long)S5P_VA_SYSRAM_NS, > >> + .pfn = __phys_to_pfn(EXYNOS5260_PA_SYSRAM_NS), > >> + .length = SZ_4K, > >> + .type = MT_DEVICE, > >> + }, > >> +}; > > As I commented before, I think we really shouldn't do this any more: There > > is no excuse why you still need to add SoC specific code here. Please put > > the SYSRAM into DT and make a proper abstraction for it so you don't have > > to modify the kernel every time a new SoC variant comes out. > > Do we have any bindings already defined for this kind of stuff or is > this implemented in > any other platform/architecture for reference? > It depends on how the sram is used. There is a generic binding in Documentation/devicetree/bindings/misc/sram.txt that may be appropriate here. Arnd -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html